faculty of chemistry , shahrood university of technology ...pollution impressed by wastewater...

2

7th Iranian Biennial Chemometrics Seminar, 30-31 October 2019

Faculty of Chemistry , Shahrood University of Technology, Shahrood, Semnan, Iran

Program book proceeding of the

7th Iranian Biennial Chemometrics

Seminar

In the name of God

30-31 Oct. 2019, Shahrood University of Technology, Shahrood, Iran

3



Welcome Message from the Seminar’s Chair

In The Name of God

It is our honor to welcome you to the 7th Iranian Biennial Chemometrics Seminar (7IBCS) at Shahrood

University of Technology on 30th and 31th October 2019. The 7IBCS provides the scientific meeting for the

Chemometricians with invited speakers and make an opportunity for younger researchers to present their

researches. The scientific program will feature to plenary talks, 11 invited speakers and 12 accepted oral

presentations and 50 posters which are distributed in two days.

It is necessary to appreciate the Shahrood University of Technology authorities; the Iranian Chemical Society,

Organizing Committee, Scientific and Referee Committees, Student Executive Committees, Novin Shimiar

Chemical Company, Fanavari Pishrafteh Jahan, Petro Kimiagar Rad and all university staffs who helped us

hold this Conference.

Sincerely Yours,

Nasser Goudarzi

Professor of Analytical Chemistry

Scientific Chair of 7IBCS

4



Executive Director: Prof. Mansour Arab Chamjangali

Shahrood University of Technology

scientific Director: Prof. Nasser Goudarzi Shahrood University of Technology

5



Dr. Hadi Parastar Sharif University of Technology

Dr. Hamid Abdollahi Institute for Advanced Studies in Basic Sciences

(IASBS), University of Zanjan

Dr. Bahram Hemmati-nezhad University of Shiraz

Scientific

Committee:

6



Dr. Abdolhossein Naseri University of Tabriz

Dr. Maryam Vosoogh Chemistry & Chemical Engineer Research Center of Iran

Dr. Mohsen Kompany-Zareh Institute for Advanced Studies in Basic Sciences (IASBS),

University of Zanjan

Dr. Morteza Bahram

University of Urumia

7



Dr. Jahanbakhsh Ghasemi

University of Tehran

Dr. Ahmad Mani

Tarbiat Modares University

Dr. Mohammad Hossein Fatemi

University of Mazandaran

Dr. Maryam Khoshkam

University of Mohaghegh Ardebili

8



Local

Organization

Committee:

Zeinab Mozafari

Ph. D. Student at

SUT

Bahare Arabkhani

Ph. D. Student at

SUT

Nahid Farzaneh

Ph. D. Student at

SUT

Farzaneh Kia

Ph. D. Student at

SUT

Amir Hossein

Momeni

Ph. D. Student at

SUT

9



Sponsors

شرکت پترو کیمیاگر راد

10



First Session

Time Program

8-8:45 Registration and Reception

8:45-9 Reading the Quran and playing the

anthem

9-9:05 Presentation of the seminar's chairman

report and greetings

9:05-9:10 Play a video clip and introducing the

University

9:10-9:20 Speech and greeting of Shahrood

University of Technology Chairman

9:20-9:30

Speech by Dr. Shamsipour, head of the

Chemical

Society of Iran

9:30-9:40 Speech by Dr. Khayamian

10:00-10:30 CoffeeBreak

Time Schedule of

7th Iranian Biannual

Chemometrics Seminar

11



Second Session: Oral Presentations

Chairmen: Dr. Naseri & Dr. Khayamian

Time Title Presenter

10:30-11:00 Do Analytical Chemists Use the

Theory of Analytical Chemistry?

Dr. Abdollahi

11:00-11:30 Class-wise LC-HRMS Data Mining

for Environmental Pollution

Monitoring

Dr. Vosoogh

11:30-11:50 Estimating confidence intervals in

multivariate curve resolution by

exploiting the principles of error

propagation in least squares

framework

Dr. Mani

11:50-12:10 How scaling can affect metabolite

identification and estimated

pathways in metabolomics studies

Dr.

Khoshkam

12:10-12:25 A new strategy for calibrating IDA-

based sensor systems

Somaiyeh

Khodadadi

Karimvand

12:25-12:40 Rapid determination of nitrate ions

in drinking water based on image

processing

techniques using a smartphone

platform

Ali Farahani

12



Third Session: Oral presentations

Chair men: Dr. Abdollahi & Dr. Fatemi

Time Title presenter

14:00-14:30

Bayesian Methods in

Chemometrics; A Simple

Introduction

Dr. Kompany-

Zareh

14:30-15:00

About the error propagation and

uncertainty estimation for the

fitted parameters using Microsoft

excel

Dr. Naseri

15:00-15:20

Some Misleading Issues in Drug

Delivery Systems and their

Associated Demands for Employing

Multivariate Chemometric

Approaches

Dr. Sajjadi

15:20-15:35

Geographical classification of olive

oil using the PLS-DA technique and

linking chemical content to classes

Mohaddeseh

rezaei

15:35-15:50

Untargeted metabolomics changes

of Gammarus Pulex in river water

induced by designed exposure with

selected pharmaceuticals: A

chemometrics study

Mahsa

Naghavi

Sheikholeslami

15:50-16:30 Coffee break and Poster presentation

13



Forth Session: Oral Prentation

Chairmen: Dr. Kompany-Zareh & Dr. Vosoogh


17:00-17:30 Where Pattern Recognition Meets

Nanostructure-Based Optical Sensors

Dr.

Hormozi-

nezhad

17:30-17:45 Essential Spectral Pixel Selection in

Hyperspectral Images

Dr.

Mahdiyeh

Ghaffari

17:45-18:00 Investigation of an interactive

molecular autoburette for simultaneous

determination of analytes by

chemometric approaches of automatic

spectrophotometric titration

Sanaz

Sajedi

Amin

City Tour 19:00-22:00

14



Fifth Session: Oral presentation

Chairmen: Dr. Goudarzi & Dr. Mani


8:45-9:15 Deep learning (past, present, future) Dr. Khosravi

9:15-9:45 Application of near infrared

spectroscopy and chemometrics for

assessing food authenticity and

adulteration

Dr. Yazdanpanah

9:45-10:05 Ensemble learning: a new concept in

chemometrics?

Dr. Parastar

10:05-10:25 Bioinformatics in drug discovery Dr. Gharaghani

10:25-11:15 Coffee break and Poster presentation

15



Sixth Session: Oral Presentations

Chairmen: Dr. Gharaghani & Dr. Prarstar


11:15-11:35 Set of Sparse Solutions in Bilinear

Decomposition

Dr. Omidikia

11:35-11:50 Convolutional neural network as a new

tool for classification of multisensor

data: prostate cancer case

Kourosh

Shariat

11:50-12:05 Application of a new hybrid of SCAD -

artificial neural network in QSAR study

of HIV inhibitors

Zeinab

Mozafari

12:05-12:20 Simultaneous determination of cysteine

enantiomers by chemometrics methods

Azam

Safarnejad

12:20-12:35

External parameter orthogonalization

combined with support vector machine

as an efficient method for analyzing

saffron NIR and ATR-FTIR spectra

to assess saffron adulteration with

plant-derived adulterants

Aryan

Amirvaresi

16



ID No. Title Page No.

OP101 A new strategy for calibrating IDA-based sensor systems

36

OP103 Rapid determination of nitrate ions in drinking water based on image processing

techniques using a smartphone platform

37

OP105 Geographical classification of olive oil using the PLS-DA technique and

linking chemical content to classes

38

OP107 Untargeted metabolomics changes of Gammarus Pulex in river water induced by

designed exposure with selected pharmaceuticals: A

chemometrics study

39

OP109 Investigation of an interactive molecular autoburette for simultaneous

40

OP111 Convolutional neural network as a new tool for classification of multisensor data: prostate cancer case

41

OP113 Application of a new hybrid of SCAD - artificial neural network in QSAR study of HIV inhibitors

42

OP115 Simultaneous determination of cysteine enantiomers by chemometrics methods

43

OP117 External parameter orthogonalization combined with support vector machine as an efficient method for

analyzing saffron NIR and ATR-FTIR spectra to

assess saffron adulteration with plant-derived

adulterants

44

Oral Presentations

17




PN001 The extraction and measurement of nickel metal ion in crab,shellfish and rice samples using magnetic silk fibroin -

EDTA ligand and chemometric method

45

PN003 Probing the binding mechanism of sorafenib to bovine α-lactalbumin using spectrometric methods, molecular docking

46

PN005 Visualization of Component-wise Rotational Ambiguity Using Signal Contribution Function

47

PN007 Local Calibration Using Multivariate Curve Resolution Methods

48

PN009 Application of Box-Behnken design and response surface methodology in optimization of salting out assisted liquid-

liquid microextraction of chromium species in environmental

samples.

49

PN011 Performance comparison of wavelet neural network and adaptive neuro -fuzzy inference system.

50

PN013 Experimental and theoretical studies on interaction of some drugs with human serum albumin.

51

PN015 The experimental and theoretical studies of Biopartitioning Micellar Chromatography to mimic the drug-

protein binding of some drugs

52

PN017 Partial least squares- residual bilinearization for simultaneous determination of ten pesticides in milk using

QuEChERS-dispersive liquid-liquid microextraction followed by

gas chromatography.

53

PN019 Random Augmented Classical Least Squares: A Modified

Calibration with CLS and ILS Advantages .

54

Poster Presentation

18




PN021 Combination of Multivariate Curve Resolution Alternating Least Squares

Method and Experimental Design to Optimize the

Simultaneous Photocatalytic Degradation of some Nitro

phenols.

55

PN023 Comparison of molecular based modelling for predicting gas heat capacity of organic compounds.

56

PN025 The feasibility of applying hand-held NIR for speciation of beef, chicken, mutton and pork with Chemometrics.

57

PN027 QSPR study of linear retention indices of some organic compounds extracted from Lupinus Pilosus Murr plants.

58

PN029 Geochemometrics Analysis of Cr, As, Hg, Cd, Pb in Tarom soil samples by spectroscopic methods.

59

PN031 The effect of random noise and spectral overlapping on the accuracy of the extracted profiles from spectroscopic data

by soft modeling method.

60

PN033 Discovery of New Inhibitors of AChE by Virtual Screening, Molecular Docking and Molecular Dynamics Simulations.

61

PN035 Multi-Stimuli Responsive Molecularly Imprinted Polymer Based on Chain Transfer Agent Modified Chitosan

Nanoparticles for Microextraction of Capecitabine: An

Experimental Design Study.

62

PN037 QSAR Study of Diarylpyrimidine Derivatives as HIV-1 Nonnucleoside Reverse Transcriptase Inhibitors by Particle

Swarm Optimization Feature Selection- Multiple Linear

Regression and Artificial Neural Networks.

63

PN039 Combining chemometrics and the TOPSIS: a new approach to optimizing HPLC parameters using multiple-responses.

64

PN041 A QSAR Study of GC-MS Retention Indecies of Essential Oils Extracted From Polygonum Minus Huds.

65

PN043 Multivariate Methods Enhanced Nontarget LC-HRMS Assessment of the River Upstream and Downstream Water

Pollution Impressed by Wastewater Treatment Effluents.

66

PN045 Quantitative structure activity relationship study of azine derivatives as NNRTIs using artificial neural network.

67

PN047 Optimization of process parameters for Paraquat and Diquat removal from binary solution by Angelica adsorbent

using Box-Behnken experimental design.

68

19




PN051 Adsorptive Removal of Phthalocyanine Using Nano-CoFe2O4 as a Sorbent from Aqueous Solution; Optimization and

Adsorption Characterization.

69

PN053 On MCR-BANDS and FACPACK under unimodality constraints.

70

PN055 Designing an IDA-based sensor array including a single indicator and receptor with multiple concentrations for

quantitation of mixtures.

71

PN059 Probing the binding mechanism of Nilotinib to bovine α-lactalbumin using

spectrometric method, molecular docking.

72

PN061 Analysis of residual moisture in a freeze-dried sample drug by multivariate fitting regression method.

73

PN063 Metabolomic study of the effects of parabens and pharmaceuticals in recycled water on metabolic pathways of

lettuce using NMR and GC-MS followed by chemometric

techniques.

74

PN065 Response surface modelling by using principal component analysis followed by partial least squares for optimizing

efficient factors in micro-solid phase extraction of polycyclic

aromatic hydrocarbons in oil spills.

75

PN067 PLS-DA vs. Q/LDA for classification of isotope ratio mass spectrometry data: a new way for food authentication.

76

PN069 Multiple response optimization of simultaneous biosorption of methylene blue and fuchsin acid by green alga Ulva

fasciata.

77

PN071 Principal component-adaptive neuro-fuzzy inference systems for the QSPR modeling of CMC of anionic gemini surfactants.

78

PN073 QSPR model for adsorption of organic compounds by multi-walled carbon nanotube (MWCNT): Comparison between MLR

and ANFIS.

79

PN075 Particle swarm optimization with various mutations for descriptor selection in QSPR studies.

80

PN077 Classification of three ground meat species using FTIR and chemometrics method.

81

PN079 Analysis of U and Th in Mahneshan soil samples by ICP-MS and Geochemometrics.

82

20




PN083 Analytical Figures of Merit for Feasible Solutions of Second-Order Calibration methods.

83

PN085 Preparation of Magnetic Molecularly Imprinted Polymer coated Multi-Walled Carbon Nanotubes for Ultra-Detection

of Sotalol: An Experimental Design Study.

84

PN087 Application Constant center and Ratio difference Methods for Simultaneous Determination of m-nitroaniline and p-

nitroaniline whit high overlapping spectra.

85

PN089 QSAR Study of New 1H-Pyrrolo [3, 2-c] Pyridine Derivatives against Melanoma Cell Lines by Firefly Algorithm-Support

Vector Machine (FF-SVM).

86

PN091 Hybrid QSPR models for the prediction of the linear retention index of volatile compounds in flour.

87

PN093 Chemometrics Study Of Dye-Surfactant Interaction By Spectroscopic And

Conductometric Methods.

88

PN095 Quantitative structure activity relationship study of DAPY-like derivatives as NNRTIs using artificial neural network.

89

PN097 Discrimination of Iranian vegetable oils by coupling of colorimetric sensor arrays and chemometrics techniques.

90

PN099 A nanozyme-based colorimetric sensor array for discrimination of anions in water

samples.

91

PN119 Multiple implementation of MARS as a new descriptor selection method in the QSAR study of a new NNRTIs using

artificial neural network

92

21



Do Analytical Chemists use "Theory of Analytical Chemistry"?

Somaiyeh Khodadadi1, Robert Rajko2, john Kalivas3 and Hamid Abdollahi1 1Department of chemistry, Institute for advanced studies in basic sciences, Zanjan, Iran

2Institute of Mathematics and Informatics, Faculty of Sciences, University of Pécs, Ifjúság

útja 6., H-7624 Pécs, Hungary 3Department of chemistry, Idaho state university, Pacatello, ID, USA

E-mail address: [email protected]

ABSTRACT

Twenty five years ago, a generalized theory of analytical chemistry (TAC) has been proposed

by Booksh and Kowalski [1]. The main point of this guiding theory is to explain about the

information and type of data which can be extracted from different analytical instrument and

methods. Accordingly, the analytical chemist can select the appropriate instrument and its

produced data based on their existing problem. Indeed, this theory can direct the analyst to

solve their research problems intelligently. The essence of theory is that, for extracting

maximum information from a determined chemical system, not only taking higher order data

from developed instruments is important, but also type of applying method is effective.

Analyzing higher order data with simple univariate methods is possible, but at the expense of

losing information.

Hence, the nature of theory is to introduce a functional framework for guiding analytical

chemist to solve their considered problems. By deep understanding of the analytical questions,

analysts should use the theory to find the optimal, practical solution. Indeed, they should design

their laboratory procedures based on the available instruments and required information to

solve the problem of interest optimally. It should be noted that, the potential and capabilities

of analytical devices and methods play a key role in choosing the optimal solutions. The figures

of merit that completely relate to the order of data, reflect the performance of the chosen

method. It can be concluded from the theory that, for optimum solving of many problems, the

analytical chemist required to take the higher order data sets, and thus apply the multivariate

methods. Also, the theory can lead analyst to design appropriate analytical devices and

methods.

Keywords: ”univariate methods”, “optimum”, “information”

References:

[1] K.S. Booksh, B.R. Kowalski, “Theory of Analytical Chemistry”, Anal. Chem., 66 (15) (1994) 782-791.

22



Class-wise LC-HRMS Data Mining for Environmental Pollution Monitoring

Maryam Vosough

Department of Clean Technologies, Chemistry and Chemical Engineering Research Center of

Iran, P.O. Box 14335-186, Tehran, Iran


ABSTRACT

The use of chromatography coupled with high-resolution mass spectrometry (HRMS) becomes

ever more important in many areas of science that rely on identification and quantification of

a large variety of compounds. Over the past few years this trend has also started in

environmental analysis, where suspect and non-target screening approaches are currently in the

focus of intense research [1]. Proper use of HRMS requires processing of “Big Data” and in

many cases insights obtained from the measured data are rather limited by inadequate data

processing and evaluation. So, developing new methodological approaches and data mining

strategies is increasingly demanding. By linking HRMS data and in-depth chemometric data

evaluation a higher level of insight into the systems under scrutiny will be achieved.

Class-wise pollution pattern studies can be considered a suitable field where multivariate

statistical and supervised classification methods can be utilized and developed. Implementation

of methods such as classic/group-wise ANOVA-simultaneous component analysis, partial least

squares-discriminant analysis and machine learning-based methods, especially in the

challenging nontarget scenarios, would prioritize the investigation of class-relevant pollutants.

Thereupon, the most meaningful pollutants that correlate with different classes of

environmental samples can be further followed. Identification and characterization of

transformation/degradation products of organic pollutants and unveiling the connections

between parent-product compounds are amongst the main benefits of these methodologies [2].

Keywords: “LC-HRMS”, “Data Mining”, “Environmental Pollution”

References: [1] J. Hollender, E. L. Schymanski, H. Singer, P.L. Ferguson, “Non-Target Screening with High Resolution Mass

Spectrometry in the Environment: Ready to Go?” Environmental Science & Technology. 51,(2017), 1505-11512.

[2] L.L. Hohrenk, M. Vosough and T.C. Schmidt, “Implementation of Chemometric Tools To Improve Data

Mining and Prioritization in LC-HRMS for Nontarget Screening of Organic Micropollutants in Complex Water

Matrixes”, Analytical chemistry, 91 (2019), 9213-9220.

23



Estimating confidence intervals in multivariate curve resolution by exploiting the

principles of error propagation in least squares framework

Ahmad Mani-Varnosfaderani 1,3*, Eun Sug Park 2, Romà Tauler3*

1 Department of Chemistry, Tarbiat Modares University, Tehran, Iran. 2 Texas A&M Transportation Institute, 3135 TAMU, College Station, TX 77843-3135, USA

3 Department of Environmental Chemistry, Institute of Environmental Assessment and Water

Research (IDAEA), Consejo Superior de Investigaciones Científicas (CSIC), Jorsi Girona 18-

25, Barcelona, 08034, Catalonia, Spain E-mail address: [email protected]

ABSTRACT

Calculation of the prediction intervals in Multivariate Curve Resolution-Alternating Least Squares (MCR-ALS) is a challenging problem. Several algorithms including Bayesian methods, Monte-Carlo

approach and jackknife resampling have been proposed previously to address this problem in MCR [1,

2]. In the present contribution, the confidence intervals (CIs) in MCR-ALS resolved profiles were estimated using the principles of error propagation in linear least squares (LS) parameter optimization.

The proposed approach is named as Confidence Intervals based on Least Squares optimization (CILS).

The weighted version of this approach has also been implemented and named as CIWLS. This method

can be used for handling datasets with a known type of error structure. The performances of the CILS and CIWLS approaches have been evaluated in this work for the estimation of the CIs for simulated

three component LC-MS and LC-DAD datasets, with different homo- and heterosedastic added noise

levels. The patterns observed for the CIs calculated using CILS and CIWLS were compared with those of Monte-Carlo noise addition and multivariate curve resolution-alternating Bayesian least square

(MCR-ABLS) approaches. The root mean squares of the differences (RMSD) between the CI5% and

CI95% values and the coverage probabilities were used as measures of the level of uncertainty in recovered profiles. The results in this work revealed that the CILS method gives similar results

compared to MCR-ABLS approach with non-informative prior for error variance. Moreover, the results

of the CILS method were in agreement with those of the Monte-Carlo approach. The main advantage

of the CILS method is that it requires less computation time and the calculations are faster. Finally, the performance of CIWLS algorithm was assessed in the analysis and source apportionment of particulate

matter (PM) air samples from a real environmental dataset collected in Northern Spain. The results

obtained by the CIWLS method were in agreement with those previously reported for this dataset.

Keywords: “error propagation”, “matrix decomposition”, “Bayesian methods”, “multivariate curve”

resolution”, “alternating least squares”, “Monte-Carlo”

References:

[1] J. Jaumot, R. Gargallo and R. Tauler, “Noise propagation and error estimations in multivariate curve resolution

alternating least squares using resampling methods”, Journal of Chemometrics, 18, (2004), 327-340.

[2] E.S. Park, M-S. Oh, Robust Bayesian multivariate receptor modeling, Chemometrics and Intelligent

Laboratory Systems, 149, (2015), 215-226.

24



How scaling can affect metabolite identification and estimated pathways in

metabolomics study

Maryam Khoshkam

Chemistry Department, Faculty of science, University of Mohaghegh Ardabili, Ardabil, Iran


ABSTRACT

Metabolomics has been successfully applied in many fields including clinical research, drug

discovery, toxicology, and phytochemistry. [1]. However extracting relevant biological

information from large data sets is a major challenge in this field [2]. From data acquisition to

statistical analysis, metabolomics data need to undergo several processing steps, which all of

them is critical in correct interpretation of data [3]. One of the most important preprocessing

method which is critical is scaling method. There has been minimal investigation of pre-

treatment methods and their influence on classification accuracy within the metabolomics

literature [4].

In this study it was observed that the reported results in metabolomics data are strongly

influenced by scaling methods. Among the methods, the effect of each method is different, and

the metabolites obtained for the same data set and the different scaling methods are quite

different. Here, a study has been conducted to investigate the influence of six pre-treatment

methods including autoscaling, range, level, Pareto and vast scaling, as well as no scaling on

three sets of 1HNMR based metabolomics data. One of the datasets was 1HNMR of mice

plasma and the other one was 1HNMR spectra of kidney and liver tissues in rattus species. The

CdTe quantum dots was injected in different doses to these animals to see the toxicity of CdTe

QDs. The results showed that in plasma and tissue data, the choice of the best scaling method

is dependent to type of datasets and in different datasets is not the same and should be checked

for each data sets. In order to investigate the best method of scaling, classification performance

parameters for each scaling method including Q2x, Q2y and R2 were computed. The resulted

metabolites and estimated biological pathways were obtained in each case. It was seen that quit

different metabolites and pathways have been obtained in each case. Thus selection of a proper

scaling methods play an important role in the metabolites identification and estimated pathway

steps in metabolomics studies.

Keywords: “phytochemistry”, “quantum dots”, “autoscaling”

References: [1] J. Yang, X. Zhao, X. Lu, X. Lin, G. Xu, “A data preprocessing strategy for metabolomics to reduce the mask effect in data analysis. Frontiers in molecular biosciences”,2, (2015), 4-12.

[2] R.A. van den Berg, H.C. Hoefsloot, J.A. Westerhuis, A.K. Smilde, M.J. “van der Werf, Centering, scaling,

and transformations: improving the biological information content of metabolomics data. BMC genomics”, 7(1),

(2012), 142-157.

[3] P.S. Gromski, Y. Xu, K. A. Hollywood, M. L. Turner, R. Goodacre, “The influence of scaling metabolomics

data on model classification accuracy”, Metabolomics, 11(3), (2015), 684-695.

[4] Craig, A., et al, Scaling and normalization effects in NMR spectroscopic metabonomic data sets. Analytical

chemistry,. 78(7), 2262-2267.

25



Bayesian Methods in Chemometrics: A Simple Introduction

Mohsen Kompany-Zareh

Department of Chemistry, Institute for Advanced Studies in Basic Sciences (IASBS), Gava

Zang, Zanjan 45137-66731, Iran. Trace Analysis Research Centre, Department of Chemistry,

Dalhousie University, PO Box 15000, Halifax, NS B3H 4R2, Canada. Email address: [email protected]

ABSTRACT

Chemometrics is increasingly being perceived as a maturing science. While this perception

seems to be true with regards to the traditional methods and applications of chemometrics.

Advances in instrumentation, computation, and statistical theory may combine to drive a

resurgence in chemometrics research. Previous surges in chemometrics research activity were

driven by the development of new ways of making better use of available information.

Bayesian statistics can further enhance the ability to use domain specific information to obtain

more accurate and useful models, and presents many research opportunities as well as

challenges.

Recent Bayesian statistical methods are based on conditional probability and practical for a

wide range of applications without making the common assumptions of Gaussian noise and

uniform prior distributions. An overview of traditional chemometric methods from a Bayesian

view and a tutorial of some recently developed techniques in Bayesian chemometrics, such as

Bayesian PCA and Bayesian latent variable regression, will be discussed. Probabilistic analysis

of non-trilinear fluorescence spectroscopic data and Naive Bayesian classification will be

considered to show the flexibility and wide range of applicability of Bayesian statistics.

Keywords: “Bayesian”, “non-trilinear”, “fluorescence spectroscopic”

References:

[1] H. Chen, B.R. Bakshi, P.K. Goel, “Toward Bayesian chemometrics-a tutorial on some recent advances”,

Anal Chim acta, 602, (2007), 1–16.

[2] K. Kumar, “Application of Akaike information criterion assisted probabilistic latent semantic analysis on

non-trilinear total synchronous fluorescence spectroscopic data sets: Automatizing fluorescence based

multicomponent mixture analysis”, Anal Chim Acta, (2019), 1062, 60-67.

[3] P. Wiczling,L. Kubik, and R. Kaliszan, “Maximum a posteriori Bayesian estimation of chromatographic

parameters by limited number of experiments”, Anal Chem, 87, (2015)7241-7249.

26



About the error propagation and uncertainty estimation for the fitted parameters using

Microsoft excel

Abdolhossein Naseri

Department of Analytical Chemistry, Faculty of Chemistry, University of Tabriz, Tabriz, Iran

E-mail: [email protected]

ABSTRACT

Uncertainty estimation and propagation of uncertainty of parameters are main topics in model

fitting. Propagation of uncertainty (or propagation of error) is the effect of variable's

uncertainties on the uncertainty of a function based on them. When the variables are the

values of experimental measurements they have uncertainties which propagate due to the

combination of variables in the function. The correlation between the variable is important

thing in error propagation which can arise from different sources. In the error propagation, if

the uncertainties are correlated then covariance must be taken into account. Unfortunately, the

correlation between variables is ignored in most chemistry textbooks to get simplicity in

calculation which leads to incorrect results [1, 2].

Microsoft Excel is the spreadsheet applications and commonly used for data analysis because

of its simplicity and universal availability [2, 3]. It has a programming ability, Visual Basic for

Applications (VBA).

The aim of this work is to show the ability of Microsoft excel in calculation of uncertainty of

parameters in fitting and also getting of their correlation in different chemical systems. Then,

propagation of uncertainty will be studied using this universal available software taking in to

account covariance matrix.

Keywords: “error propagation”, “correlation”, “covariance matrix”, “fitting”, “Microsoft excel”

References:

[1] J. N. Miller and J. C. Miller, “Statistics and Chemometrics for Analytical Chemistry”, 5th Edition, 2005,

Pearson Education Limited.

[2] R. De Levie, Advanced Excel for Scientific Data Analysis, second Edition, 2019.

[3] A. Naseri, H. Khalilzadeh and S. Sheykhizadeh, “Tutorial Review: Simulation of Oscillating Chemical

Reactions Using Microsoft Excel Macros”, Analytical and Bioanalytical Chemistry Research, 3(2), (2016), 169-

185.

mailto:[email protected]://en.wikipedia.org/wiki/Variable_(mathematics)https://en.wikipedia.org/wiki/Uncertaintyhttps://en.wikipedia.org/wiki/Function_(mathematics)https://en.wikipedia.org/wiki/Observational_errorhttps://en.wikipedia.org/wiki/Correlatedhttps://en.wikipedia.org/wiki/Covariance

27



Some Misleading Issues in Drug Delivery Systems and their Associated Demands for

Employing Multivariate Chemometrics Approaches

S. Maryam Sajjadi

Faculty of Chemistry, Semnan University, Semnan, Iran. E-mail address: [email protected]

ABSTRACT

Drug delivery systems (DDSs) refer to the pharmaceutical technology employed for presenting

the drug to the desired body site. In therapeutic goals, it is an urgent need to control the delivery

of drugs in both desired dose and site, in order to decrease their adverse side effects. In the

study of DDSs, pharmacokinetic investigations have gained much attention from researchers;

however, in this regard, there are some misleading issues such as decomposition of some drugs

during their release process [1]. Indeed, any changes in DDSs’ formulations, either

quantitatively or qualitatively, could influence drug release; therefore, it is crucial to have a

deep insight into the mechanism of drug release and its side reactions.

There are a verity kind of multivariate chemometrics methods that can be utilized to find the

kinetic mechanisms and estimate the profiles of all or some kinetic profiles of species involved

in the reactions [2]. It should be noted that high-performance liquid chromatography (HPLC)

is a common method used for most of DDSs assessments because it can produce selective

responses for drug [3]. However, there is no limitation of using spectrophotometric methods

for monitoring the kinetic processes in DDSs when they are coupled with chemometrics

strategies which are able to resolve the data containing spectroscopically active species with

severely overlapped signals.

In this study, it will be discussed how chemometrics approaches can be applied successfully to

investigate different kinetic processes in drug delivery systems which are responsive to light,

pH, or temperature. Moreover, it will be shown how the loading condition of drug can influence

its release mechanism.

Keywords: “Drug Delivery”, “Pharmaco-kinetic”, “Multivariate Data; “Chemometric Methods”

References:

[1] H. Etezadi, SM. Sajjadi, A. Maleki, “New Journal of Chemistry”, 43 (2019), 5077-5087.

[2] R. Tauler, B.Walczak, SD. Brown, Comprehensive Chemometrics. Chemical and Biochemical Data

Analysis. Elsevier, 2009.

[3] B. Mao, C. Liu, W. Zheng, X. Li, R. Ge, H. Shen, X. Guo, Q. Lian, X. Shen and C. Li, “Biomaterials”, 161,

(2018), 306-320.

28



Where Pattern Recognition Meets Nanostructure-based Optical Sensors

M Reza Hormozi-Nezhad

Department of Chemistry, Sharif University of Technology, Tehran, Iran


ABSTRACT

Visual detection, as a universal sensing approach, holds great promise in various fields such as

environmental monitoring, food safety, security issues, clinical and point-of-care diagnosis,

and healthcare assays in especially resource-constrained areas, where sophisticated

instrumentation may not be available. Many efforts have been made over the past few decades

to develop optical probes which can be classified into two main groups: colorimetric and

fluorometric approaches. In the former, changes in the absorption signal or wavelength is

assigned to the concentration of an analyte whereas in the latter, changes in the emission

characteristics is monitored for quantification. In either of these analytical signal modes,

implementation of nanostructures can greatly enhance sensing. Owing to the high extinction

coefficient of plasmonic nanoparticles together with their size, shape and environment

dependent absorption profiles, they provide much better colorimetric probes compared to their

conventional counterparts. Similarly, it has been shown that emitters in nanoscale such as

quantum dots, nanoclusters and metal organic framework materials with incredible and tunable

emission properties, have recently attracted great attention in the fields of sensing and

bioimaging. Moving from single optical probes towards cross-reactive sensor arrays enables

the recognition and discrimination of groups of target species. In array-based sensors, instead

of relying on a specific lock and key interaction for sensing, an array of semi-selective sensor

elements is used. These cross-reactive sensor elements provide differential responses and

generate measurable fingerprint patterns which are analyzed by pattern recognition methods in

order to classify the data and to detect unknown samples. Since the large amount of data

provided in a sensor array usually has a high dimension and cannot be analyzed by basic

calibration methods, a multivariate data reduction method is required to reduce the dimension

of the data and to make it better visually interpretable [1-2].

In this presentation, basic principles in the design of nanostructure-based optical sensor arrays

will be outlined. Focusing on our recent research in this field [3], we will present several

examples of luminescent and plasmonic nanoparticles that have been used to produce the

desired assembly of sensor elements for detection and discrimination of important analytes. Keywords: “Bayesian”, “non-trilinear”, “fluorescence spectroscopic”

References:

[1] J. R. Askim, M. Mahmoudi, K. S. Suslick “Optical sensor arrays for chemical sensing: the optoelectronic

nose”, Chemical Society Reviews, 42(22), (2013), 8649-8682.

[2] A. Bigdeli, F. Ghasemi, H. Golmohammadi, S. Abbasi-Moayed, M. A. F. Nejad, N. Fahimi-Kashani, M.

Shahrajabian, M. R. Hormozi-Nezhad “Nanoparticle-based optical sensor arrays”, Nanoscale, 9(43), (2017),

16546-16563.

[3] F. Ghasemi, M. R. Hormozi-Nezhad, M. Mahmoudi “A colorimetric sensor array for detection and

discrimination of biothiols based on aggregation of gold nanoparticles”, Analytica chimica acta, 882, (2015), 58-

67.

mailto:[email protected]

29



Essential Spectral Pixel Selection in Hyperspectral Images

Mahdiyeh Ghaffari1, Nematollah Omidikia2, Cyril Ruckebusch1

1Univ. Lille, CNRS, UMR 8516 LASIR, F-59000 Lille, France, 2Department of Chemistry, Faculty of Science, University of Sistan and Baluchestan,

P .O. Box 98135-674, Zahedan, Iran.


ABSTRACT

Spectral imaging techniques are now important tools in chemical analysis. These tools combine

the spectroscopic attributes of chemical measurements with the ones of imaging. Challenging

applications of spectral imaging can be found in chemistry, biology, medicine, food science or

agriculture, at both the micro and macro scales [1,2]. However, while chemical images can now

be acquired in routine, linear spectral unmixing with multivariate curve resolution (MCR),

which assumes a low-rank approximation of the bilinear decomposition to extract spectra of

the pure/est individual chemical components and distribution of their proportions in the image

of a scene, remains a challenging problem. Despite recent advances both in the field of self-

modeling curve resolution (SMCR) and on the practical side, (bio) chemical are difficult to

analyze, because they are big and spatial-spectral information is highly mixed [3]. In this work

we propose a methodology to select Essential Spectral Pixels (ESPs) (very important pixel

instead) of chemical images. These pixels are on the outer envelope of the principal component

scores of the data and can be identified by convex-hull computation. As they carry all the

spectral information that is useful for linear unmixing, all other measured pixels can be

removed resulting in simpler multivariate curve resolution (MCR) analysis of large

hyperspectral images. The proposed procedure is used to analyze several chemical images of

different spectroscopies, sizes and complexities and show that multivariate curve resolution

analysis done on full data sets of hundreds of thousands of spectral pixels can be performed on

reduced data sets composed of very sparse sets of ESPs.

Keywords: “Essential Spectral Pixels (very important pixels)”, “Multivariate Curve Resolution”, “convex-hull”, “SMCR”, “chemical images”

References:

[1] T.V. Galassi, P.V. Jena, D. Roxbury, D.A. Heller, “Single nanotube spectral imaging to determine molar concentrations of isolated carbon nanotube species”, Analytical chemistry, 89, (2017), 1073-1077. [2] H.J. Butler, L. Ashton, B. Bird, G. Cinque, K. Curtis, J. Dorney, K. Esmonde-White and M.J. Walsh, , “Using Raman spectroscopy to characterize biological materials”, Nature protocols, (2016), 664–687. [3] B. Prats-Mateu, M. Felhofer and A. Juan, “Multivariate unmixing approaches on Raman images of plant cell walls: new insights or overinterpretation of results?”, Plant Methods, 14, (2018), 52.

30



Deep learning (past, present, future)

Hossein Khosravi

Faculty of Electrical and Robotic Engineering, Shahrood University of Technology,

Shahrood, P.O. Box 36199-95161, Iran.


ABSTRACT

Deep learning is making a big impact in many areas of human life for solving complex

problems. Deep learning models share various properties and the learning dynamics of neurons

in human brain. It covers many areas of artificial intelligence, including image classification,

image captioning, machine translation, speech recognition, drug discovery and computational

chemistry.

The main concept of deep learning is not new, it is about 30 years old. With the development

of large data sets, huge computing power and new algorithms, the true power of the concepts

are now revealed.

In this lecture we will review the historical perspective of deep learning including:

Perceptron, the first model of neural network

Backpropagation and MLP

First Deep Network introduced by LeCun 1989

Recurrent Neural Networks

Restricted Boltzman Machine

ImageNet and its Influence on development of Deep Learning

GPUs and their Influence on Deep Learning Furthermore, we will describe the present status of deep learning including:

Convolutional Networks

Regional Proposal Networks (R-CNN)

Deep Recurrent Networks (RNN, LSTM)

Deep Reinforcement Learning (Q-Learning)

Applications of DNN

And finally, a few things about the future directions for deep learning:

Quantum deep learning

Automated Machine Learning

Competing Learning Models

Hybrid Learning Models

Keywords: “Deep learning”, “Convolutional Networks”, “Quantum deep learning”

31



Application of near infrared spectroscopy and chemometrics for assessing food

authenticity and adulteration

Hassan Yazdanpanah

Food Safety Research Center, Shahid Beheshti University of Medical Sciences, Tehran, IR

Iran E-mail address: [email protected]

ABSTRACT

Food authenticity and adulteration are major issues in the food industry and are attractive for

consumers. The globalization of our food makes it more vulnerable to food adulteration with

both unintentional and intentional fraud being perpetrated. The latter is very often used for

economic gain, also called “economically motivated adulteration”. The risk for food

adulteration increases proportionately with the complexity of the supply chain. Considering the

fact that the prediction of possible adulterants is not always an easy or sometimes even a

possible task, this in turn leads to opportunities to make huge financial gains with a very low

risk of detection. This can make the task of deciding which analytical methods are more

suitable to collect and analyse chemical data within complex food supply chains, at targeted

points of vulnerability, that much more challenging. It is evident that those working within and

associated with the food industry are seeking rapid, user-friendly methods to evaluate food

authenticity and adulteration, and rapid/high-throughput screening methods for the analysis of

food in general. In addition to being robust and reproducible, these methods should be portable

and ideally handheld and/or remote sensor devices, that can be taken to or be positioned on/at-

line at points of vulnerability along complex food supply networks and require a minimum

amount of background training to acquire information rich data rapidly (ergo point and-shoot).

There are several methods available to characterize authenticity of foods, but these methods

are usually expensive and time consuming. In relation to the globally traded amount of foods,

an adequate number of controls by several traditional analytical methods is not realistic. In

contrast, near infrared (NIR) as a spectroscopic fingerprinting technique has been shown to be

a low cost, rapid, convenient, precise, multi-analytical and non-destructive screening method

for food authentication and adulteration. Along with chemometrics, a resolution of unique

chemical information is provided, which allows rapid monitoring of subtle compositional

changes. Therefore, the comparison of the fingerprints obtained from authentic samples to

adulterated samples can reveal mis-description or adulterations. The use of NIR as an analytical

tool for process control, food safety and quality has been well recognized and accompanied by

the application of chemometrics for data pre-treatment and analysis and multivariate screening

and modelling. The NIR as a rapid method could be networked and thus used to detect trends

in the food market perhaps even before any food security threat/event is acknowledged by

regulators and thus could very easily sit within the umbrella of the Internet of Things. A big

advantage of spectrometric methods combined with chemometrics lies in the fact that once a

database is established and a suitable data analysis protocol is determined, a new sample can

be screened within a few minutes. With a suitable user interface, even non-specialist personnel

(such as food inspectors and consumers) can undertake sample analysis on-site as well as in

QC laboratories and factories.

Among foods, meat and fruit juices are among commodities that meet the criteria for a high

risk of being affected by adulteration. In this regard, we evaluated the feasibility of a handheld NIR device (900 – 1700 nm) for speciation of mutton, beef, chicken, and pork. NIR

32



spectroscopy was coupled with two different chemometric methods including Partial Least

Squares Discriminant Analysis (PLS-DA) and Support Vector Machine (SVM). After spectral

acquisition, the 6 spectra of each sample was used for further analysis. Spectral datasets were

divided into calibration (70%) and validation (30%) sets with duplex algorithm and pre-

processed with Mean Center and 2nd derivative (Savitzky–Golay) for PLS-DA and SVM-C

models. The best results were achieved with SVM model. For SVM model, sensitivity and

specificity values in the validation set were 88% and 94% for mutton, 95% and 99% for beef,

84% and 96% for chicken, and 86% and 93% for pork, respectively. SVM model overall

accuracy was 87%. The finding presents, for the first time, the potential of hand-held NIR

spectroscopy with chemometrics models for rapid, inexpensive and non-destructive speciation

of 4 different types of raw meat samples.

In another study, we investigated the novel application of a handheld NIR device (900 – 1700

nm) coupled with multivariate classification methodologies as a screening approach in

detection of adulterated lime juices. Three diffuse reflectance spectra of 31 pure lime juices

(collected from Jahrom, IR. Iran) and 25 adulterated ones were acquired. Principal component

analysis was almost able to generate two clusters. PLS-DA and k-nearest neighbors algorithms

with different spectral preprocessing techniques were applied as predictive models. In the PLS-

DA, the most accurate prediction was obtained with SNV transforming. The generated model

was able to classify juices with an accuracy of 88% and the Matthew’s correlation coefficient

value of 0.75 in the external validation set. In the k-NN model, the highest accuracy and

Matthew’s correlation coefficient in the external validation set (88% and 0.76, respectively)

was obtained with multiplicative signal correction followed by 2nd-order derivative and 5th

nearest neighbor.

The results showed that handheld NIR in combination with multivariate analysis can be a very

promising rapid first-step screening method for evaluation of meat and lime juice authenticity.

Handheld NIR is, therefore, an ideal tool for high throughput analysis of a high number of

samples identifying suspects which require further examination by state-of-the-art

confirmatory methods.

Keywords: “Chemometrics”, “Near Infrared”, “Food authenticity”, “Food adulteration”, “Lime juice”, “Meat

species”

33



Ensemble learning: a new concept in chemometrics?

Hadi Parastar

Department of Chemistry, Sharif University of Technology, Tehran, Iran E-mail address: [email protected]

ABSTRACT

Ensemble learning is a machine learning paradigm where multiple learners are trained to

solve the same problem. In contrast to ordinary machine learning approaches which try to

learn one hypothesis from training data, ensemble methods try to construct a set of

hypotheses and combine them to use [1]. An ensemble contains a number of learners which

are usually called base learners. The generalization ability of an ensemble is usually much

stronger than that of base learners. Actually, ensemble learning is appealing because that it is

able to boost weak learners which are slightly better than random guess to strong learners

which can make very accurate predictions. So, “base learners” are also referred as “weak

learners” [2].

Ensemble learning is a new concept in computer science for analysis of “Big Data” especially

for classification and regression purposes [3]. However, the potential use of this method is

under question in chemometrics. Therefore, in this contribution, the concept of ensemble

learning and different types of learners is discussed and then their potential for the analysis

of chemical data is investigated. Furthermore, its performance will be compared with

conventional chemometric methods. As an example, random subspace discriminant ensemble

(RSDE) [4] as one of ensemble learning algorithms combined with handheld near-infrared

(NIR) spectroscopy is used to show the potential of ensemble learning for analysis of chemical

data. In this regard, we developed a powerful method to test chicken meat authenticity. The

research presented in this work shows that it is both possible to discriminate fresh from thawed

meat, based on NIR spectra, but even to correctly classify chicken fillets according to the

growth conditions of the chickens with good accuracy. In all cases, the RSDE method

outperformed other common classification methods such as partial least squares-discriminant

analysis (PLS-DA), artificial neural network (ANN) and support vector machine (SVM) with

classification accuracy of >95%. This study shows that handheld NIR coupled with machine

learning algorithms is a useful, fast, non-destructive tool to identify the authenticity of chicken

meat. By comparing and combining different protocols to measure the NIR spectra (i.e.,

through packaging and directly on meat), we show the possibilities for both consumers and

food inspection authorities to check the authenticity of packaged chicken fillet. Keywords: “Ensemble learning”, “Chemometrics”, “Classification; Machine learning”

References: [1] L. Rokach, Artificial Intelligence Review, 33, )2010(,1-39.

[2] C. Merkwirth, H. Mauser, T. Schulz-Gasen, O. Roche, M. Stahl, T. Lengauer, "Ensemble methods for

classification in cheminformatics.", Journal of Chemical Information and Computer Sciences, 44(6), (2004),

1971-1978.

[3] H. Parastar, R. Tauler, Angewandte Chemie: International Edition, (2019), xx, xxx-xxx.

[4] T. K. Ho, IEEE Transactions on Pattern Analysis and Machine Intelligence, 20, (1998), 832-844.

34



Bioinformatics in drug discovery

Sajjad Gharaghani*

Laboratory of Bioinformatics and Drug Design, Institute of Biochemistry and Biophysics,

University of Tehran, Tehran, Iran E-mail address: [email protected]

ABSTRACT

The human community still faces the problem of finding an effective and potent drug. Until

now, the discovery of new drugs has not been in line with the advancement of science and

technology. The methods of drug design are divided into ligand-based and structure-based

categories. Ligand-based methods include Quantitative Structure Activity Relationship

(QSAR) and pharmacophore models. In the structure-based approach (molecular docking), the

protein-drug interaction is usually used for modeling. While this method leads to the discovery

of the active compounds, it fails in the clinical phase due to the side effect. Many of the side

effects are due to drug interactions with off-target proteins. Therefore, the need for

computational methods that take into account drug interactions with all proteins seems

essential. Nowadays, using bioinformatics methods with machine learning and network-based

approach considers drug and protein interaction networks to provide a solution to this problem.

Keywords: “Drug discovery”, “Bioinformatics”, “QSAR, “Docking”, “Pharmacophore”

References:

[1] Medina-Franco, J.L, Giulianotti, Marc A, Welmaker, Gregory S., Houghten, Richard A, "Shifting from the

single to the multitarget paradigm in drug discovery", Drug Discovery Today, 18, (2013), 495-501.

35



Set of Sparse Solutions in Bilinear Decomposition

Nematollah Omidikia*

Department of Chemistry, Faculty of Science, University of Sistan and Baluchestan,

P .O. Box 98135-674, Zahedan, Iran


ABSTRACT

Several constraints are designed to restrict bilinear decompositions further to get a unique

solution [1]. Sparsity constraint was introduced to create solutions with zero elements [2]. As

nor the number of zeros neither the place of zeros are not initially available, sparsity constraint

should be incorporated with caution. Regarding sparsity constraint, two important issues can

be addressed. The first issue is the effect of sparsity constraint on the possible solutions of

bilinear decompositions, finding the set of sparse solutions. The second issue is the type of Lp-

norm, {p=0,1,2}, for the sparsity implementation. Focusing on the geometry of bilinear data

sets, outer-polygon as the non-negativity boundary in curve resolution contains all the possible

sparse solutions [3]. In this contribution, we shed light on the all possible sparse solutions, and

it was shown that outer-polygon is the set of sparse solutions. Not only sparse solution, but also

sparset solutions are located on the outer boundaries. Finally, Lp-norms were calculated for the

different feasible profiles, and it is revealed that L0 minimization and L2 maximization are

correct strategies to reach the sparse/est solutions. However, L1-norm is not appriate candidate

for sparse non-negative decomposition.

Keywords: “unique solution”, “bilinear data sets”, “feasible profiles”

References: [1] N. Omidikia, H. Abdollahi, and M. Kompany-Zareh,, “On uniqueness and selectivity in three-component

parallel factor analysis”, Analytica chimica acta, 782, (2013), 12–20.

[2] M. Ghaffari, S. Hugelier, L. Duponchel, H. Abdollahi, and C. Ruckebusch, “Effect of image processing constraints on the extent of rotational ambiguity in MCR-ALS of hyperspectral images”, Analytica Chimica Acta,

1052, (2019), 27-36.

[3 R. Rajkó and K. István, “Analytical solution for determining feasible regions of self‐modeling curve resolution (SMCR) method based on computational geometry”, Journal of the Chemometrics Society, 19.8, (2005), 448–

463.

36



A new strategy for calibrating IDA-based sensor systems

Somaiyeh Khodadadi Karimvand, Hamid Abdollahi *,

Department of Chemistry, Institute for Advanced Studies in Basic Sciences, Zanjan, Iran E-mail address: [email protected]

ABSTRACT

Nowadays, the indicator displacement assay (IDA) has wide applications in chemistry due to

its high potential to selective and sensitive determination of different analytes [1]. In order to

calibrating IDA-based sensor arrays systems for simultaneous quantification of analytes,

nonlinear calibration methods such as Artificial Neural Network (ANN) are mainly used. For

minimizing the complication of the constructed models in different algorithms of ANN

methods and hence for preventing the over fitting problem, removing of redundant input

variables is unavoidable. For this reason, various variable selection methods have been

introduced to performing this important task. So, as the results of calibration model in ANN

are completely affected by the chosen variables, herein, we present a novel strategy for

calibrating the IDA-based colorimetric sensors with ANN models using dramatically reduced

number of input variables and without needing any variable selection method. The general

unique feature of IDA-based sensor systems is that the species of signal generators, the

indicator and probe (indicator-receptor complex), are known and their pure spectra can be

easily available. As the target analyte(s) in IDA sensors is colorless, any obtained data from

these systems are the result of changing in the equilibrium concentrations of these two species.

Herein, we proposed that for calibrating IDA-based sensors, instead of using signals with a

large number of variables, the equilibrium concentration of active species with smaller number

of variables can be replaced. As a result, the number of input variables in the calibration and

thus, the possibility of overfitting will be significantly reduced. Most equilibrium chemical systems including IDA sensors due to presence of matrix effect are intrinsically non-linear. So,

the Beer-Lambert law is not valid in the non-linear systems, and the simple least square method

(CLS) of data to pure spectrum does not result in correct free concentrations profiles. Thus, in

this situations the generalized classical least square or Indirect Hard Modelling (IHM) approach

can be applied as an alternative method for resolving the equilibrium concentrations of the

spectroscopic active species [2]. The performance of the proposed strategy was evaluated in a

designed sensor array for simultaneous quantification of Histidine and Cysteine.

Keywords: "Indicator Displacement Assay (IDA)", "Sensor array", "Simultaneous quantification", "Artificial

Neural Network (ANN)", "Indirect Hard Modelling (IHM)", "Histidine and Cysteine"

References: [1] B. T. Nguyen and E. V. Anslyn, “Indicator Displacement Assay”, Coordination Chemistry Reviews, 250,

(2006), 3118-3127.

[2] F. Alsmeyer and HJ. Koß, W. Marquardt, “Indirect spectral hard modelling for the analysis of reactive and

interacting mixtures”, Appl Spectrosc, 58, (2004), 975‐985.

https://www.sciencedirect.com/science/article/abs/pii/S0010854506001147#!https://www.sciencedirect.com/science/article/abs/pii/S0010854506001147#!https://www.sciencedirect.com/science/journal/00108545

37



Rapid determination of nitrate ions in drinking water based on image processing

techniques using a smartphone platform

Ali Farahani, Hassan Sereshti*, Camelia Tashakori, Shamim Azimi

School of Chemistry, College of Science, University of Tehran, Tehran, Iran. E-mail address: [email protected]

ABSTRACT

Developing a simple operating, robust and affordable technique for monitoring the

concentration of nitrate residues in drinking water has become a crucial issue which has led to

the introduction of various nitrate determination methods [1]. The present study aims to

introduce a rapid, low-cost and portable smartphone-platform sensing device based on image

processing techniques for fast determination of nitrate ions in water samples. A sample holder

and photography kit, as shown in Fig.1A, was designed using cost-effective components. A

circle of 600 pixels in diameter with the best correlation and sensitivity to nitrate ion

concentration, was chosen as the region of interest using Convolutional Neural Network (CNN)

shown in Fig.1B. An application platform named nitrate hunter developed and lunched in

smartphone by app inventor platform (Fig.1C). Application was conducted to measure the

nitrate levels in drinking water samples collected from 42 different zones of Tehran and Alborz

provinces (Iran). The nitrate level in each sample was determined using a smartphone and UV-

vis spectrophotometry. Based on the statistical analysis, nitrate concentration read from UV-

visible spectrometer and that of calculated from the smartphone provided high correlation of

R2=0.982. Correspondingly, using image processing and deep learning techniques nitrate

concentration were detected in the range of 5 to 100 µg mL-1 (Fig.1D) with a high

determination coefficient (R2 = 0.995). Besides, the smartphone device predicted an LOQ

value (5.0 µg mL) lower than the maximum residual level set by WHO.

Figure 1: procedure of test and validation nitrate level in drinking water.

Keywords: “Nitrate ions”, “Smartphone-based techniques”, “Water quality determination”, “Image processing”

References:

[1] G. Hua, M. W. Salo, C. G. Schmit and C. H. Hay, “Nitrate and phosphate removal from agricultural subsurface

drainage using laboratory woodchip bioreactors and recycled steel byproduct filters”, Water Research, 102,

(2016), 180–189.

38



Geographical classification of olive oil using the PLS-DA technique and linking

chemical content to classes

Mohaddeseh rezaei, Mohsen Kompany Zareh*, Maryam Khoshkam

Institute for Advanced Studies in Basic Sciences


ABSTRACT

One of the most important issues of the olive oil industry in Iran is the definition of the grade

of olive oil. Internationally, most countries use the standards of the International Association

of Olive Oils (IOOC) to define the quality of olive oil. Due to the high nutritional value of olive

oil and its benefits to human health, there are many traditional methods for defining the grade

of olive oil, but none of these methods are sufficient on their own and thus the definition of an

economical, comprehensive and simple method for the definition of the quality of olive oil is

important. In this study, the least squares split strain analyzer (PLS-DA) method and method

(Multi-Block Data Analysis) has been used as a method for determining the relationship

between the quality of olive oil. This study is important because in Iran, there is no way to

classify olive oils according to their nutritional value and their geographical area, so most of

the counterfeit olive oils are sold instead of virgin olive oil.

In addition to the study done in comparison with similar work in the world, it is advantageous

to combine and compare the results of the PLS-DA and Multi-Block Data Analysis methods.

The results of previous studies have shown that the issue of non-compliance of the quality label

on glass of olive oil with its actual quality is a serious issue in Iran. In this research, it will be

shown that these methods can be used as a comprehensive method for defining the degree of

quality of olive oil successfully. This method can be an easy and economical way to define the

grade of olive oil quality for the olive oil extraction industry [1-3].

Keywords: “olive oil”, “PLS-DA”, “Multi-Block Data Analysis”, “geographical”, “classification”, “chemical content”

mailto:[email protected]

39



Untargeted metabolomics changes of Gammarus Pulex in river water induced by

designed exposure with selected pharmaceuticals: A chemometrics study

Mahsa Naghavi Sheikholeslami1, Maryam Vosough*,1 , Roma Tauler*2 , Cristian Gomez-

Canela2 1Department of Clean Technologies, Chemistry and Chemical Engineering Research Center

of Iran, Tehran, Iran, P.O. Box 14335-186 Tehran, Iran 2 Department of Environmental Chemistry, Institute of Environmental Assessment and Water

Research (IDAEA), Consejo Superior de Investigaciones Científicas (CSIC), Barcelona,

08034, Catalonia, Spain E-mail address: [email protected]

ABSTRACT

Recently, the presence of pharmaceuticals and personal care products (PPCPs) in water due to

incomplete removal in wastewater treatment plants (WWTPs) is a serious concern. In this work

the effect of three pharmaceuticals (Propranolol, Triclosan, and Nimesulide) exposure on

Gammarus pulex metabolic profiles in river water was assessed by liquid chromatography

coupled to high resolution mass spectrometry (LC-HRMS), in an untargeted way [1]. The

generated complex data sets in the different exposure experiments were processed by different

chemometric tools based on the selection of regions of interest (ROIs) and on multivariate

curve-resolution alternating least squares (MCR-ALS). Utilizing analysis of variance

simultaneous component analysis (ASCA) on metabolite peak profile areas resolved by MCR-

ALS showed significant changes between different contaminants, different pharmaceutical

concentrations (exposed and non-exposed samples) and between different exposure times (2h,

6h and 24h). In addition, 34 metabolites were common between various contaminants, which

they have been interpreted using ASCA [2,3].

Keywords: “PPCPs”, “WWTPs”,“Gammarus pulex”, “Metabolite”, “LCHRMS”, “ROI-MCR-ALS”, “ASCA”

References: [1] C. Gómez-Canela, T.H. Miller, N.R. Bury, R. Tauler, and L.P. Barron, “Targeted metabolomics of Gammarus

pulex following controlled exposures to selected pharmaceuticals in water”, Science of The Total Environment,

562, (2016), 777-788.

[2] C. Gómez-Canela, E. Prats, B. Piña and R. Tauler, “Assessment of chlorpyrifos toxic effects in zebrafish

(Danio rerio) metabolism”, Environmental pollution, 220, (2017), 1231-1243. [3] E. Gorrochategui, J. Jaumot, and R. Tauler, “A protocol for LC-MS metabolomic data processing using

chemometric tools”, Protocol Exchange, 2015.

40



Investigation of an interactive molecular autoburette for simultaneous

determination of analytes by chemometric approaches of automatic

spectrophotometric titration

Sanaz Sajedi Amin1, Abdolhossein Naseri1, Hamid Abdollahi*,2 1 Department of Analytical Chemistry, Faculty of Chemistry, University of Tabriz, Tabriz,

Iran 2 Department of Chemistry, Institute for Advanced Studies in Basic Sciences, Zanjan, Iran


ABSTRACT

The device is a tool that invented or constructed for a special goal. Chemical devices are

molecular level devices, which can be used instead of physical ones and also expected to

open the way to revolutionizing the science i.e. in drug delivery and solving the environmental

pollution [1].A chemical system provides to change the value of an nenvironmental parameter

(pH, temperature, etc.) inside a reaction vessel in a controlled

way without any interfering with the progression of studied reaction. Here we studied the

possibility of using any chemical devices that provide variable pH condition as a molecular

burette in reaction vessel based on model based analysis. The chemical compounds such as

cryptand, tert-buthylchloride or any chemical system that produce or entrap H+ in reaction

vessel can act as a variable pH autoburette [2]. Proper simulation of this mechanism based on

the kinetic or intertwined equilibrium-kinetic model; enable ones to design an experimental

direction for its use as a molecular burette. The present study aims to investigate the optimum

condition of molecular outoburette operating parameters to obtain better performance for

various acid-base titration. In other words, a larger pH range of action or a balancing the rate

of pH change, provide the almost ideal demand system for automatic titrations. So, the effect

of different factors such as initial concentration of chemical device, starting pH and buffer

capacity on tuning the pH-time profile of molecular burette were investigated. Finally, the

proposed molecular device can be evaluated for simultaneous determination of binary mixtures

of food colorants by chemometrics analysis of simulated pectrophotometric titration data,

which was alternative to traditional extensive series of experiments. This kind of data structure,

analyzed by multivariate curve resolution-alternating least squares (MCR-ALS) under the non-

negativity, correspondence and trilinearity constraints [3] As a result, the concentration of each

dye in the samples and their corresponding pure spectra were obtained. Keywords: “molecular devices”, “automatic titration”, “food colorants”, “multivariate curve resolution-

alternating least squares”

References:

[1] V. Balzani, A. Credi, and M. Venturi, “Molecular devices and machines: concepts and perspectives for

the nanoworld”, John Wiley & Sons, 2008.

[2] G. Alibrandi and Cryptandm “A Chemical Device for Variable‐pH Kinetic Experiments”,Angewandte Chemie International Edition, 47, (2008), 3026-3028.

[3] H.C. Goicoechea, A.C. Olivieri, and R. Tauler, “Application of the correlation constrained multivariate

curve resolution alternating least-squares method for analyte quantitation in the presence of

unexpected interferences using first-order instrumental data”, Analyst, 135, (2010), 636-642.

41



Convolutional neural network as a new tool for classification of multisensor data:

prostate cancer case

Kourosh Shariat1, Dimitry Kirsanov2, Hadi Parastar*,1 1Department of Chemistry, Sharif University of Technology, Tehran, Iran

2Departmetnt of Analytical Chemistry, Saint Petersburg State University, Saint Petersburg,

Russia


ABSTRACT

Convolutional neural networks (CNNs) have shown excellent performance in the past few

years on a variety of machine learning problems to process multidimensional data and to

recognize local patterns which makes them useful for problems such as image analysis and

sound recognition [1]. Additionally, a recent study showed that CNNs can be efficiently applied

to classify vibrational spectroscopic data and it was claimed that CNN outperformed

conventional classification methods such as partial least squares-discriminant analysis (PLS-

DA), support vector machine (SVM) and logistic regression (LR) [2]. This implies the

relevance of feasibility study of CNNs as a possible tool for data analysis in other applications.

Prostate cancer (PCa) is the second most common cancer in males and it is one of the leading

causes of cancer mortality. Early detection of prostate cancer is crucial for successful therapy

and so far, the common methods of detection are either inaccurate or resource-consuming [3].

The potentiometric multisensor systems are the arrays of cross-sensitive electrodes which can

be used for untargeted detection of biomarkers and therefore, a multivariate “fingerprint” can

be obtained. The main objective of the present contribution was development of a chemometric

classification method based on CNN for classification of multisensor data of prostate cancer

towards early diagnosis of this cancer. The studied data set contained 89 samples (43 from

biopsy confirmed PCa patients and 46 from control group) characterized with responses from

28 sensors [3]. The original data were splitted into calibration and test set using duplex

algorithm. Then, different preprocessing methods including mean-centering, auto-scaling and

smoothing were tested on the performance of CNN. In optimum CNN condition, CNN gave

95.6% sensitivity, 97.7% specificity and 95.0% accuracy which were really surprising. Also,

CNN performance was superior of the conventional classification methods of PLS-DA and

SVM. It is concluded that CNNs can be effectively used to classify multisensor data and more

importantly, detect prostate cancer via potentiometric multisensor systems at early stage.

Keywords: “Convolutional neural networks”, “Chemometrics”, “Multisensor system”, “Prostate cancer”

References:

[1] J. Gu, Z. Wang, J. Kuen, L. Ma, A. Shahroudy, B. Shuai, B. Liu, X. Wang, L. Wang, G. Wang, J. Cai and T. Chen, “Recent Advances in Convolutional Neural Networks”, arXiv, 1, (2015), 1512-07108.

[2] J. Acquarelli, T. van Laarhoven, J. Gerretzen, L. M.C. Buydens and E. Marchiori, “Convolutional neural networks for vibrational spectroscopic data analysis”, Analytica Chimica Acta, 954, (2017), 22-31.

[3] S. Solovieva, M. Karnaukh, V. Panchuk, E. Andreev, L. Kartsova, E. Bessonova, A. Legin, P. Wang, H. Wan,

I. Jahatspanian and D. Kirsanov, “Potentiometric multisensor system as a possible simple tool for non-invasive prostate cancer diagnostics through urine analysis”, Sensors and Actuators B: Chemical , 289, (2019), 42-47.

42



Application of a new hybrid of SCAD - artificial neural network in QSAR study of HIV

inhibitors

Zeinab Mozafari*, 1, Mansour Arab Chamjangali1, Mohammad Arashi2, Nasser Goudarzi 1

1Department of Chemistry, Shahrood University of Technology, Shahrood, P.O. Box 36155-

316, Iran. 2Department of Statistics, Faculty of Mathematical Sciences, Shahrood University of

Technology, Shahrood, P.O. Box: 316-3619995161, Iran. E-mail address: [email protected]

ABSTRACT

A hybrid of smoothly clipped absolute deviation (SCAD) and Levenberg- Marquardt (LM)

artificial neural network (ANN) was used as a new approach in the quantitative structure-

activity relationship (QSAR) studies. Fan and Li presented the SCAD in 2001 to improve

previous variable selection methods’ [1] performance. The SCAD has advantages such as

unbiased estimation, continuity, low prediction error, stability, good sparsity and high

interpretation. Hence, SCAD as an oracle method has an efficient penalty function. Recently,

SCAD has been used as modeling method in QSAR/QSPR studies [2,3].

57 new HIV inhibitors were used in the QSAR modeling and pEC50 of compounds were

simulated. 3224 Dragon descriptors were computed for the thioacetamide/acetanilide

derivatives [4]. Dataset were divided into the three categories of train set (35 compounds),

validation set (11 compounds) and the test set (11 compounds). SCAD method [2] was applied

on the train and validation set data (46 compounds) and 11 non-zero coefficients corresponded

to the parameter with the lowest cross validation error (λmin) were selected and used as inputs

of modeling method. The predictability of the optimum LM-ANN model was evaluated using

external test, leave one out (LOO) technique and statistical parameters. Statistical parameters

such as determination coefficient (R2) and mean square error (MSE) of the test set were 0.92

and 0.12 respectively, which prove the generalizability and predictability of the constructed

model. According to the effects of descriptors on the biological activity, some active

compounds were suggested and the interaction of ligand-enzyme were analyzed using

Autodock4.2 and pyMOL softwares.

Keywords: “HIV”, “QSAR”, “SCAD”, “Artificial neural network”, “Molecular docking”

References: [1] J. Fan, R. Li, “Variable selection via nonconcave penalized likelihood and its oracle properties”, Journal of

the American statistical Association, 96 (2001) 1348-1360.

[2] Z.Y. Algamal, M.H. Lee, A novel molecular descriptor selection method in QSAR classification model based

on weighted penalized logistic regression, Journal of Chemometrics, 31 (2017) e2915. [3] Z.Y. Algamal, M.H. Lee, A.M. Al‐Fakih, High‐dimensional quantitative structure–activity relationship

modeling of influenza neuraminidase a/PR/8/34 (H1N1) inhibitors based on a two‐stage adaptive penalized rank

regression, Journal of Chemometrics, 30 (2016) 50-57.

[4] X. Li, X. Lu, W. Chen, H. Liu, P. Zhan, C. Pannecouque, J. Balzarini, E. De Clercq, X. Liu, “Arylazolyl

(azinyl) thioacetanilides. Part 16: Structure-based bioisosterism design, synthesis and biological evaluation of

novel pyrimidinylthioacetanilides as potent HIV-1 inhibitors”, Bioorganic & medicinal chemistry, 22 (2014) 5290-5297.

43



Simultaneous determination of cysteine enantiomers by chemometrics methods

Azam Safarnejad1, M. Reza Hormozi-Nezhad2,3, Hamid Abdollahi *1

1 Department of Chemistry, Institute for Advanced Studies in Basic Sciences, P.O. Box 45195-

1159, Zanjan, Iran Zanjan, Iran 2 Department of Chemistry, Sharif University of Technology, Tehran, 11155-9516, Iran.

3 Institute for Nanoscience and Nanotechnology, Sharif University of Technology, Tehran,

Iran E-mail address: [email protected]

ABSTRACT

The determination and analysis of chiral compounds are of critical importance in chemical and

pharmaceutical sciences. The Cysteine amino acid is one of the important chiral compounds

that each enantiomer (L and D) has different effects on fundamental physiological Processes.

Th

faculty of chemistry , shahrood university of technology ...pollution impressed by wastewater...

Documents